Quantifying Thermo-Mechanical Stability in Binder Jetting Cores: A Statistical Approach with Coating and Orientation Effects

Abstract Dimensional stability and thermo-mechanical performance are critical aspects of any core manufacturing process. Among additive manufacturing (AM) technologies, binder jetting is increasingly adopted in sand casting; understanding how process parameters affect these properties becomes particularly important. This study investigates the effect of printing configurations (i.e., part orientation and position inside the build box) and post-process steps (i.e., baking and coating) on the thermo-mechanical behaviour of furan binder-jetted silica mix sand cores using the Hot Distortion Test (HDT). Results show that core baking alone induces only modest modifications (≈ 20% increase in deflection and slope of the HDT curve, but negligible impact on printing anisotropy), whereas the application of a coating (with subsequent baking) substantially enhances thermal durability and extends the breakage time by ≈ 40% on average without increasing maximum deflection. More importantly, coating mitigates the strong anisotropic behaviour of vertically printed specimens, reducing the performance gap with planar orientations, with deflection differences dropping from ~ 25 to ~ 12% after coating. Multivariate regression analysis consistently highlights the printing orientation of the parts and post-process as the main significant factors across both experiments, while the part location inside the printer build box has only a minor or inconsistent effect. To improve the analysis, an HDT-area descriptor is designed and adopted, which aggregates multiple curve features into a single metric. Regression models based on HDT-area confirm the same significant factors while providing excellent explanatory power (Adj- R 2 up to 85%) and reducing printing orientation-dependent variability. These findings indicate that coating not only prolongs core durability but also reduces part-to-part variability, while HDT-area emerges as a reliable synthetic indicator for design and quality assurance in AM-enabled foundry workflows.